Make code DRYer, and make it possible for this to work under Py2 and Py3.

gsnedders · gsnedders · commit 70922940dd6b · 2013-01-08T23:05:53.000Z
diff --git a/html5lib/treebuilders/dom.py b/html5lib/treebuilders/dom.py
@@ -1,28 +1,12 @@
 
 from xml.dom import minidom, Node, XML_NAMESPACE, XMLNS_NAMESPACE
-try:
-    from types import ModuleType
-except:
-    from new import module as ModuleType
 import re
 import weakref
 
 from . import _base
 from html5lib import constants, ihatexml
 from html5lib.constants import namespaces
-
-moduleCache = {}
-
-def getDomModule(DomImplementation):
-    name = "_" + DomImplementation.__name__+"builder"
-    if name in moduleCache:
-        return moduleCache[name]
-    else:
-        mod = ModuleType(name)
-        objs = getDomBuilder(DomImplementation)
-        mod.__dict__.update(objs)
-        moduleCache[name] = mod    
-        return mod
+from html5lib.utils import moduleFactoryFactory
 
 def getDomBuilder(DomImplementation):
     Dom = DomImplementation
@@ -285,6 +269,11 @@ def dom2sax(node, handler, nsmap={'xml':XML_NAMESPACE}):
         
     return locals()
 
+
+# The actual means to get a module!
+getDomModule = moduleFactoryFactory(getDomBuilder)
+
+
 # Keep backwards compatibility with things that directly load 
 # classes/functions from this module
 for key, value in list(getDomModule(minidom).__dict__.items()):
diff --git a/html5lib/treebuilders/etree.py b/html5lib/treebuilders/etree.py
@@ -1,30 +1,13 @@
-try:
-    from types import ModuleType
-except:
-    from new import module as ModuleType
 import re
-import types
 
 from . import _base
 from html5lib import ihatexml
 from html5lib import constants
 from html5lib.constants import namespaces
+from html5lib.utils import moduleFactoryFactory
 
 tag_regexp = re.compile("{([^}]*)}(.*)")
 
-moduleCache = {}
-
-def getETreeModule(ElementTreeImplementation, fullTree=False):
-    name = "_" + ElementTreeImplementation.__name__+"builder"
-    if name in moduleCache:
-        return moduleCache[name]
-    else:
-        mod = ModuleType("_" + ElementTreeImplementation.__name__+"builder")
-        objs = getETreeBuilder(ElementTreeImplementation, fullTree)
-        mod.__dict__.update(objs)
-        moduleCache[name] = mod    
-        return mod
-
 def getETreeBuilder(ElementTreeImplementation, fullTree=False):
     ElementTree = ElementTreeImplementation
     class Element(_base.Node):
@@ -342,3 +325,6 @@ def getFragment(self):
             return _base.TreeBuilder.getFragment(self)._element
         
     return locals()
+
+
+getETreeModule = moduleFactoryFactory(getETreeBuilder)
diff --git a/html5lib/treewalkers/__init__.py b/html5lib/treewalkers/__init__.py
@@ -8,6 +8,8 @@
 returning an iterator generating tokens.
 """
 
+from importlib import import_module
+
 treeWalkerCache = {}
 
 def getTreeWalker(treeType, implementation=None, **kwargs):
@@ -34,7 +36,7 @@ def getTreeWalker(treeType, implementation=None, **kwargs):
     treeType = treeType.lower()
     if treeType not in treeWalkerCache:
         if treeType in ("dom", "pulldom", "simpletree"):
-            mod = __import__(treeType, globals())
+            mod = import_module("."+treeType, "html5lib.treewalkers")
             treeWalkerCache[treeType] = mod.TreeWalker
         elif treeType == "genshi":
             from . import genshistream
diff --git a/html5lib/treewalkers/etree.py b/html5lib/treewalkers/etree.py
@@ -1,31 +1,15 @@
 import gettext
 _ = gettext.gettext
 
-try:
-    from types import ModuleType
-except:
-    from new import module as ModuleType
 import copy
 import re
 
 from . import _base
 from html5lib.constants import voidElements
+from html5lib.utils import moduleFactorFactory
 
 tag_regexp = re.compile("{([^}]*)}(.*)")
 
-moduleCache = {}
-
-def getETreeModule(ElementTreeImplementation):
-    name = "_" + ElementTreeImplementation.__name__+"builder"
-    if name in moduleCache:
-        return moduleCache[name]
-    else:
-        mod = ModuleType("_" + ElementTreeImplementation.__name__+"builder")
-        objs = getETreeBuilder(ElementTreeImplementation)
-        mod.__dict__.update(objs)
-        moduleCache[name] = mod
-        return mod
-
 def getETreeBuilder(ElementTreeImplementation):
     ElementTree = ElementTreeImplementation
 
@@ -139,3 +123,5 @@ def getParentNode(self, node):
                     return parent, list(parents[-1]).index(parent), parents, None
 
     return locals()
+
+getETreeModule = moduleFactoryFactory(getETreeBuilder)
diff --git a/html5lib/utils.py b/html5lib/utils.py
@@ -5,6 +5,9 @@
     from sets import Set as set
     from sets import ImmutableSet as frozenset
 
+from types import ModuleType
+
+
 class MethodDispatcher(dict):
     """Dict with 2 special properties:
 
@@ -173,3 +176,25 @@ def surrogatePairToCodepoint(data):
     char_val = (0x10000 + (ord(data[0]) - 0xD800) * 0x400 + 
                 (ord(data[1]) - 0xDC00))
     return char_val
+
+# Module Factory Factory (no, this isn't Java, I know)
+# Here to stop this being duplicated all over the place.
+
+def moduleFactoryFactory(factory):
+    moduleCache = {}
+    def moduleFactory(baseModule, *args, **kwargs):
+        if type(ModuleType.__name__) is str:
+            name = "_%s_factory" % baseModule.__name__
+        else:
+            name = b"_%s_factory" % baseModule.__name__
+
+        if name in moduleCache:
+            return moduleCache[name]
+        else:
+            mod = ModuleType(name)
+            objs = factory(baseModule, *args, **kwargs)
+            mod.__dict__.update(objs)
+            moduleCache[name] = mod    
+            return mod
+
+    return moduleFactory
